EP0183877A1 - Mikrorechner für zeitabhängige Prozesse - Google Patents

Mikrorechner für zeitabhängige Prozesse Download PDF

Info

Publication number
EP0183877A1
EP0183877A1 EP84308344A EP84308344A EP0183877A1 EP 0183877 A1 EP0183877 A1 EP 0183877A1 EP 84308344 A EP84308344 A EP 84308344A EP 84308344 A EP84308344 A EP 84308344A EP 0183877 A1 EP0183877 A1 EP 0183877A1
Authority
EP
European Patent Office
Prior art keywords
timer
time
processes
processor
collection
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
EP84308344A
Other languages
English (en)
French (fr)
Other versions
EP0183877B1 (de
Inventor
Michael David May
Roger Mark Shepherd
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Inmos Ltd
Original Assignee
Inmos Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Inmos Ltd filed Critical Inmos Ltd
Priority to DE8484308344T priority Critical patent/DE3483007D1/de
Priority to EP84308344A priority patent/EP0183877B1/de
Priority to KR1019850006994A priority patent/KR930009757B1/ko
Publication of EP0183877A1 publication Critical patent/EP0183877A1/de
Application granted granted Critical
Publication of EP0183877B1 publication Critical patent/EP0183877B1/de
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/48—Program initiating; Program switching, e.g. by interrupt
    • G06F9/4806—Task transfer initiation or dispatching
    • G06F9/4812—Task transfer initiation or dispatching by interrupt, e.g. masked
    • G06F9/4825—Interrupt from clock, e.g. time of day

Definitions

  • the invention relates to computers including microcomputers and is particularly applicable to microcomputers capable of executing time dependent processes.
  • a microcomputer is described in our European Patent Specification 0110642 which includes scheduling means to permit the processor to share its processing time between a plurality of concurrent processes.
  • a linked list of scheduled processes awaiting execution may be formed.
  • a currently executing process may be descheduled and processes may be scheduled by adding to a scheduled list when required. This may for example arise in effecting message transmission between two processes where it is required that both processes be at corresponding stages in their program sequence when the message transmission occurs.
  • patent specification does not describe the use of time dependent processes wherein scheduling of a process may be effected in accordance with a specfified time for the process.
  • microcomputer relates to small size computers generally based on integrated circuit devices but it does not impose any limit on how small a computer may be.
  • the present invention provides a microcomputer comprising memory and a processor arranged to execute a plurality of concurrent processes each in accordance with a program consisting of a plurality of instructions for sequential execution by the processor, said processor comprising (1) a plurality of registers and data transfer means for use in data transfers to and from said registers (2) means for receiving each instruction and loading into one of the processor registers a value associated with the instruction, and (3) control means for controlling said data transfer means and registers in response to each instruction received to cause the processor to operate in accordance with the instruction, wherein the microcomputer includes:-
  • Preferably means are arranged to order the processes in said timer collection to form a time ordered sequence in dependence on the scheduling time of the or each process in the collection.
  • said memory provides for each process a workspace having a plurality of addressable locations including locations for recording variables associated with the process, and in which one of said processor registers is arranged to hold a workspace pointer value identifying an address of the workspace of the current process.
  • the workspace for each process includes link means for holding a pointer value for a subsequent process in the timer collection, said link means being used when a process is in the timer collection to indicate the next process in the collection thereby forming a linked timer list of processes.
  • said link means of each process workspace is arranged to hold a special value to indicate that the process is currently the process with the last scheduling time on the timer list.
  • the workspace for each process includes an addressable location for indicating the scheduling time of the process.
  • the microcomputer includes means for allocating one of a plurality of priorities to each process and said time control means include means for forming more than one said timer collection, each timer collection having processes of a common priority which is different from that of processes in another timer collection.
  • the scheduling means includes:
  • the scheduled collection is also a linked list.
  • the invention also includes a network having a plurality of interconnected microcomputers as aforesaid, each having communication channels provided by one or more communication links which are connected by dedicated connections to similar links on further devices thereby permitting message transmission with synchronisation between concurrent processes on different microcomputers.
  • the microcomputer is arranged to execute a process with a plurality of alternative time related components, said microcomputer being provided with means to indicate a time associated with each component, means to test the time associated with each component and means to determine whether the earliest time associated with a component has yet occurred.
  • Preferably means are provided for specifying a time duration for the execution of a process and means responsive to said time duration to cause the processor to stop executing the current process after expiry of the time duration and to reschedule the process by adding it to a scheduled collection.
  • the microcomputer described in this example comprises an integrated circuit device in the form of a single silicon chip having both a processor and memory in the form of RAM as well as links to permit external communication.
  • the main elements of the microcomputer are illustrated in Figure 1 on a single silicon chip 11 using p-well complementary MOS technology.
  • a central processing unit (CPU) 12 is provided with a timer 9 to allow time control of the execution of processes. It also includes some read-only memory (ROM) 13 and is coupled to a memory interface 14 controlled by interface control logic 15.
  • the CPU 12 incorporates an arithmetic logic unit (ALU), registers and data paths illustrated more fully in Figure 2.
  • the CPU 12 and memory interface 14 are connected to a bus 16 which provides interconnection between the elements on the chip 11.
  • a service system 17 is provided with a plurality of input pins 18.
  • the microcomputer is provided with a random access memory (RAM) 19 and ROM 20 and the amount of memory on the chip is not less than 1 K byte so that the processor 12 can be operated without external memory.
  • the memory on the chip is at least 4 K bytes.
  • An external memory interface 23 is provided and connected to a plurality of pins 24 for connection to an optional external memory.
  • a plurality of serial links 25 are provided having input and output pins 26 and 27 respectively.
  • the input and output pins of one serial link may each be connected by its own single wire non-shared unidirectional connection to the corresponding output and input pins of a serial link on another microcomputer as shown in Figure 8.
  • Each serial link is connected to a synchronisation logic unit 10 comprising process scheduling logic.
  • the present embodiment provides an improved form of Transputer (Trade Mark of INMOS International plc) microcomputer. It provides for timer control so that processes may be executed in dependence on timer data and timer lists of processes awaiting specified times before execution may be formed.
  • Transputer Trade Mark of INMOS International plc
  • the overall arrangement of the microcomputer is generally similar to that described in the above mentioned patent applications. In the following description similar names will be given to those parts corresponding to the embodiment in the above mentioned patent applications.
  • the memory provides a plurality of process workspaces having addressable locations which can be indicated by pointers. Message communication can be effected through channels which may comprise addressable memory locations in the case of process to process communication on the same microcomputer. To effect process to process communication between different microcomputers input and output channels are provided in serial links and these channels may also be addressed in a manner similar to the locations provided in the memory.
  • the particular- wordlength of the example described is 16 bits but it will be understood that other wordlengths such as 8, 16, 24, 32 or other wordlengths may be used.
  • different wordlength microcomputers can be connected in the same network as shown in Figure 8 so that they may communicate with each other regardless of their independent wordlength.
  • Each pointer is a single word and is treated as a two's complement signed value. That means that if the most significant bit in the pointer is a 1 the most signficant bit is taken as negative with all the remaining bits representing positive numbers. If the most significant bit is 0 then all bits in the pointer are taken as representing positive values. This enables the standard comparison functions to be used on pointer values in the same way that they are used on numerical values.
  • TimeSet.p and TimeNotSet.p are never used in the same locations as Enabling.p or Waiting.p so that no ambiguity arises from the dual use of the values MostNeg + 1 and MostNeg + 2.
  • each process has a workspace consisting of a vector of words in memory used to hold the local variables and temporary values manipulated by the process.
  • a workspace pointer WPTR is used to point to a set location for the process workspace.
  • Each process can be identified by a "process descriptor" of which the least significant bit indicates the priority of the process and the most significant 15 bits indicate the word in memory identifying the process workspace.
  • the microcomputer allocates one of two possible priorities to each process.
  • each process descriptor comprises a single word which is formed by taking the "bitwise OR" of the workspace pointer WPTR and the process priority Pri.
  • the workspace pointer WPTR can be obtained from a process descriptor by forming the "bitwise AND” of the process descriptor and NOT 1.
  • the priority of the process can be obtained by forming the "bitwise AND” of the process descriptor and 1.
  • the CPU 12 includes an arithmetic logic unit (ALU) 30 and a plurality of data registers connected to an X bus, Y bus, Z bus and bidirectional data bus 31.
  • ALU arithmetic logic unit
  • the operation of the registers and their interconnections with the buses is controlled by a plurality of switches diagrammatically represented at 32 and controlled by signals derived from a microinstruction program contained in the ROM 13. Communication between the CPU and the memory is effected via a unidirectional address path 33 leading to the memory interface 14 as well as the data bus 31.
  • each instruction consists of 8 bits, 4 bits representing the required function of the instruction and 4 bits being allocated for data.
  • Each instruction derived from the program sequence for the process is fed to an instruction buffer 34 and the instruction is decoded by a decoder 35.
  • the output of the decoder is fed through a condition multiplexor 36 to a microinstruction register 37 used for addressing the microinstruction ROM 13.
  • the operation of the instruction buffer 34, decoder 35, condition multiplexor 36, MIR 37, microinstruction ROM 13 and switches 32 are generally as described in the above mentioned patent applications, and in European Patent Specification 0110642.
  • Register bank 38 is provided for the priority 1 processes and a similar register bank 39 is provided for the high priority 0 processes. Both register banks have a similar set of registers similarly connected to the X, Y, Z and data buses. For simplicity, the registers and their connections have only been shown in detail for register bank 38.
  • the CPU includes a constants box 40, a register bank selector 41 and a number of other registers indicated in Figures 2A and 2B which are common-to both priority 0 and priority 1 processes.
  • the registers are as follows:-
  • the bank of registers 39 for priority 0 processes is the same as that already described for priority 1 processes.
  • the suffix [1] indicates a register relevant to the priority 1 bank and the suffix [0] indicates that the register relates to the priority 0 bank.
  • the suffix [Pri] indicates that a register of appropriate priority to the process is used.
  • the registers are generally of word length which in this case is 16 bits apart from the 1 bit flags 47, 48, 58, 59, 82, 83 and 84.
  • the instruction buffer may be of 8 bit length if arranged to hold only 1 instruction at a time.
  • the A, B and C register stack 54, 55 and 56 are the sources and destinations for most arithmetic and logical operations. They are organised as a stack.
  • each of the banks 38 and 39 includes TIMER LOGIC 86 arranged to receive inputs from the VALID TIME FLAG 84, the next TIME REG 85 and the CLOCK REG 81.
  • the TIMER LOGIC 86 will be described more fully with reference to Figure 3.
  • the CLOCK REG 81 receives an input from a PROCESSOR CLOCK 87.
  • the TIMER LOGIC 86 for each of the register banks constitutes the timer 9 of Figure 1.
  • the OREG 57 of both register banks 38 and 39 are connected to the decoder 35 so that for both priority processes that part of the instruction which is fed into the OREG register reaches the decoder for use in generating appropriate microinstructions.
  • the SNP FLAG 58, COPY FLAG 59, INSERT FLAG 82, DELETE FLAG 83 and TIMER LOGIC 86 of both priority banks are also connected to the condition multiplexor 36 so that the microinstructions can take into account the setting of these flags and the logic output for either priority process in determining the next action to be effected by the processor at any time.
  • the constants box 40 is connected to the Y bus and enables constant values to be placed on that bus under the control of the microinstruction ROM 13. These can be used in pointing to offset locations in a process workspace and providing time slice periods.
  • the register bank selector 41 has inputs from the PRI FLAG 47, the PROCPRI FLAG 48 and the microinstruction ROM 13. The output from the register bank selector is connected to the condition multiplexor 36, to the decoder 35 and to the switches 32. Depending on the output of the microinstruction ROM 13, the selector will chose the register bank indicated by the PRI FLAG 47 or the PROCPRI FLAG 48.
  • the TIMER LOGIC 86 is similar for each of the register banks and one is shown more fully in Figure 3.
  • the logic unit 86 comprises a subtractor 88 arranged to receive an input on line 89 from the NEXT TIME REG and this time value is subtracted in the subtractor 88 from the time value supplied on a line 90 from the CLOCK REG 81.
  • the most significant bit of the difference is provided on an output on line 91 to an inverter 92 which supplies a signal on line 93 to as logical AND gate 94.
  • the gate 94 also receives an input on line 95 from the VALID TIME FLAG 84.
  • the AND gate 94 provides an output on line 96 which is fed to the condition multiplexor 36.
  • the signal on line 96 is called a "Timer Request” signal and is arranged to cause the processor to remove a process from the top of a timer list so that it becomes ready for execution. This will be described more fully below. It will be appreciated that the logic diagram shown in Figure 3 is arranged so that a "Timer Request" signal on line 96 is only output when two conditions are met simultaneously. Firstly the VALID TIME FLAG 84 must be set to the value 1 and the time indicated by the CLOCK REG 81 must either be after or equal to the time indicated by the NEXT TIME REG 85.
  • the subtractor 88 is used to subtract the value contained in the NEXT TIME REG 85 from the value held in the CLOCK REG 81 and if the result of that subtraction is a negative number the most significant bit will be 1 due to the use of two's complement signed values as referred to above. For this reason line 91 is arranged to output the most significant bit resulting from the subtraction and the inverter 92 is required so that the AND gate 94 only provides the "Timer Request" when the result of the subtraction provides a positive result thereby causing a 0 bit on line 91.
  • the microcomputer carries out a number of processes together sharing its time between them. Processes which are carried out together are called concurrent processes and at any one time the process which is being executed is called the current process. Each concurrent process has a region of memory called a workspace for holding the local variables and temporary values manipulated by the process.
  • the address of the first local variable of the workspace is indicated by the workspace pointer (WPTR).
  • WPTR workspace pointer
  • FIG 4 This is indicated in Figure 4 where four concurrent processes, Processes L, M, N and 0 have workspaces 60, 61, 62 and 63.
  • the workspace 60 has been shown in more detail and the workspace pointer held in the WPTR REG 51 points to the 0 location which is a single word location having the address indicated in this example as 10000.
  • the other local variables for this process are addressed as positive offset addresses from the word indicated by the workspace pointer.
  • Some of the workspace locations with small negative offsets from the 0 location are used for scheduling timing and communication purposes.
  • five additional word locations 65, 66, 67, 68 and 69 are shown having negative offsets of 1, 2, 3, 4 and 5 respectively below the 0 location indicated by the WPTR. These locations are as follows:-
  • Location 65 is used when a process is not the current process to hold a pointer (IPTR) to the next instruction to be executed by the process when it becomes the current process.
  • Location 66 is used to store a workspace pointer of a next process on a linked list or queue of scheduled processes awaiting execution.
  • Location 67 is normally used to contain an indication of the state of a process performing an alternative input operation or as a pointer for copying of a block of- data.
  • Location 68 is used to store a workspace pointer of a next process on a linked timer list of processes awaiting predetermined times before being scheduled for execution and it is also used to indicate the state of a process performing an alternative timer input operation.
  • Location 69 is used to indicate a time after which the process may be executed.
  • the memory also provides word locations for process to process communication and Figure 3 indicates such a channel 70.
  • Procedure "StartNextProcess" deschedules the current process and, if there is another runnable process, selects the next runnable process. This may cause the resumption of an interrupted priority 1 process if there are no further priority 0 processes to run.
  • Tptr LocO In the above definition reference is made to Tptr LocO. It will be appreciated that there are two Tptr locations, one for priority 1 and another for priority 0. They occupy adjacent memory locations and that for priority 0 has the address TptrLocO. In this way either of the locations can be addressed by an offset of 0 or 1 from TptrLocO depending on the relevant priority.
  • InsertStep is executed as a result of the InsertFlag[Pri] being set. Repeated performance of this procedure will insert the current process into the timer list for the current priority level in the correct position.
  • the Breg[Pri] and Creg[Pri] registers identify the point at which the search for the correct location has so far reached.
  • the procedure "DeleteStep” is executed as a result of the DeleteFlag[Pri] being set. Repeated performance of this procedure will delete the current process from the timer list for the current priority level.
  • the Breg[Pri] and Creg[Pri] registers identify the point at which the search for the current process has so far reached.
  • the processor performs a sequence of actions. These are performed either on behalf of the current process, or on behalf of a request made by a serial link 25 or the timer 9.
  • a “priority 1 action” is correspondingly defined.
  • the actions which may be performed on behalf of the current process are the procedures "StartNextProcess”, “InsertStep”, “DeleteStep”, “BlockCopyStep” or to fetch, decode and execute an instruction.
  • Each of these actions corresponds to a sequence of microinstructions.
  • the last microinstruction in any of the sequences comprising these actions is "NextAction". This causes the processor to choose the next action to be performed.
  • the way in which the processor decides which action is to be performed next when a "NextAction" microinstruction is executed is as follows.
  • the sync control logic 10 will forward to the processor at most one "RunRequest” or "ReadyRequest” at any time. It will not forward a priority 1 request if there is a priority 0 request outstanding. This results in two signals being input to the condition multiplexor 36, one indicating the presence of a request and the other indicating the priority of that request.
  • the two signals "TimerRequest0" and “TimerRequestl” are connected to the condition multiplexor 36 which is also connected to signals from the currently selected SNPFlag 58, DeleteFlag 83, InsertFlag 82 and CopyFlag 59. It is therefore able to make the selection as described below.
  • the processor will perform the procedure "StartNextProcess” if the SNPFlag[Pri] is set. Otherwise the processor will select a priority 0 action is there is one that can be performed. Otherwise the processor will select a priority 1 action if there is one that can be performed. Otherwise the processor will wait until there is a request from a timer or communication channel.
  • the processor selects an action at a particular priority level Pri according to the following rules.
  • the processor will perform a "DeleteStep” if the DeleteFlag[Pri] is set. Otherwise the processor will perform an "InsertStep” if the InsertFlag[Pri] is set. Otherwise the processor will handle any priority Pri channel request. Otherwise the processor will handle any priority Pri timer request. Otherwise the processor will perform the procedure "BlockCopyStep" if the CopyFlag[Pri] is set. Otherwise the processor will fetch, decode and execute an instruction if there is a current process of priority Pri.
  • the CLOCK REG 81 increments by 1 regularly and goes through continuous cycles incrementing from the most negative value up to the most positive value. The next increment after the most positive value takes the register back to the most negative value.
  • the expression (X AFTER Y) means X is later than the time Y. All times between (X + 1) and (X + MostPos) are defined to be AFTER X. All times which are between (Clock Reg + 1) and (Clock Reg + MostPos) are considered to be in the future and those which are between (Clock Reg and (- 1)) and (Clock Reg + MostNeg) are considered to be in the past.
  • each instruction for the microcomputer includes a function element selected from a function set.
  • the functions executed by the microcomputer include direct functions, the prefixing functions pfix and nfix, and an indirect function opr which uses the operand register Oreg to select one of a set of operations.
  • Oreg[Pri] is cleared after the execution of all instructions except PFIX and NFIX.
  • microinstruction ROM 13 contains microinstructions corresponding to all the above listed functions and operations whereby the processor is caused to carry out any of the above actions as a result of microinstructions derived from the ROM 13.
  • the processor shares its time between a number of concurrent processes executing at the two different priority levels 0 and 1.
  • a priority 0 process will always execute in preference to a priority 1 process if both are able to execute.
  • WPTR workspace pointer
  • IPTR instruction pointer
  • Such a scheduled list is formed as a linked list with each process on the list having a pointer 'in the link location 66 of its workspace to the workspace of the next process on that list.
  • the instruction pointer (IPTR) of any process on the list is stored in the IPTR location 65 of its workspace as shown in Figure 4.
  • the processor may maintain two lists of scheduled -processes which are waiting to be executed, one for each priority level.
  • it may maintain two timer lists of descheduled processes awaiting specified times before being scheduled, one timer list being provided for each priority.
  • Figure 4 indicates the high priority 0 scheduled list whereas Figure 5 shows a low priority 1 scheduled list at a time when a priority 0 process is the current process as shown in Figure 4.
  • the register bank selector 41 has selected the registers in bank 39 for use by the processor. Consequently WPTR REG [0] holds a pointer to the 0 location of the workspace 60 of the current process L as indicated in Figure 4.
  • the IPTR REG [0] contains a pointer to the next instruction in the program sequence 181 which is stored in memory.
  • the registers 54, 55, 56 and 57 indicated in Figure 4 contain other values to be used during execution of the current process L.
  • the scheduled list of priority 0 processes which are awaiting execution is indicated in Figure 4 by the three processes M, N and 0 whose workspaces are indicated diagrammatically at 61, 62 and 63. Each of these workspaces is generally similar to that indicated for process L.
  • the FPTR REG [0] marked 53 contains the pointer to the workspace of process M which is the process at the front of this list.
  • the workspace of process M contains in its IPTR location 65 a pointer to the next instruction in the program sequence which is to be executed when process M becomes the current process.
  • the link location 66 of process M contains a pointer to the workspace of process N which is the next process on the list.
  • the last process on the list indicated is process 0 which has its workspace indicated at 63.
  • the BPTR REG [0] marked 52 contains a pointer to the workspace of this last process 0.
  • the workspace 63 of this process 0 is pointed to by the contents of the link location 66 of the previous process N but in this case the link location 66 of process 0 does not contain any pointer as this is the last process on the list.
  • a pointer to the workspace of that further process is placed in the BPTR REG 52 and the link location 66 of the process 0 then contains a pointer to the workspace of the further process which is added to the list.
  • the priority 1 scheduled list is generally similar and this is indicated in Figure 5.
  • the list of priority 1 processes which have been scheduled and are awaiting execution consists of the processes P, Q and R.
  • a further priority 1 process marked S is shown but this is currently descheduled and does not form part of the linked list.
  • the FPTR REG [1] contains a pointer to the workspace of process P which forms the first process on the list awaiting execution.
  • the BPTR REG [1] contains a pointer to the workspace of process R which forms the last process on the scheduled list.
  • Each of the processes P, Q and R has an IPTR in its IPTR location pointing to the program stage from which the next instruction is to be taken when that process becomes the current process.
  • the link location of each process apart from the last process on the scheduled list contains a pointer to the workspace of the next process on the list.
  • a process may be taken from the top of a list for execution by use of the procedure "dequeue" which has been defined already.
  • a current process may be descheduled by the procedure "start next process" which has been defined already.
  • the present embodiment does however provide a time slicing facility such that if the current process is a low priority process it may be stopped after a period of time called a "time slice" and rescheduled at the end of the queue illustrated in Figure 5 so as to allow the opportunity for other processes on the scheduled list to be executed.
  • the processor executes the procedure "dequeue" and as can be seen from lines 11 and 12 of the definition of that procedure, if the process is a priority 1 process (which will be the case for a low priority process) then according to line 12, the TIME SLICE REG 80 is loaded with a value which is the sum of the present time indicated by the CLOCK REG 81 together with the time required "length of time slice".
  • the length of a time slice may be chosen to suit any appropriate time interval and in the present case it is taken to be the time needed to execute 1000 instructions. This may of course be varied as necessary. This time slice is stored in the constants box 40.
  • the processor When the low priority process executes a "jump" function or a "loop end” operation, the processor carries out the procedure "time slice” as can be seen from the end of the definition of both the jump function and the loop end operation.
  • the processor checks that if the priority of the current process is 1 and the time indicated by the CLOCK REG is equal to or after the time indicated by the TIME SLICE REG 80 then the sequence is carried out in which the workspace pointer and priority of the current process is loaded into the PROC DESC REG 46 and the procedure "run” is carried out so that the process is rescheduled by adding it to the end of the priority 1 scheduled list.
  • the procedure also sets the SNPFlag 58 to the value 1 so that the processor ceases executing the current process and starts to execute a further process from the top of the priority 1 scheduled list unless there is any process or request of higher priority requiring action by the processor.
  • the present embodiment also makes provision for timer lists of the type shown in Figures 6 and 7.
  • Figure 6 illustrates a linked timer list of low priority 1 processes whereas Figure 7 shows a similar linked list of high priority 0 processes.
  • the low priority processes have been marked in Figure 6 with the letters T, U and V whereas the high priority processes of Figure 7 have been given the letters W, X and Y.
  • the two lists are generally similar and for this reason only the list of Figure 6 will be described in detail.
  • the workspace 60 for each of the processes in the list is indicated in Figure 6.
  • the front of the timer list is maintained by a single word memory location 90 which holds a pointer value called TPTR.
  • the TPTR for that priority is set to the special value "NotProcess.p". Otherwise the TPTR held in the memory location 90 points to the "variable 0" location (also called 0 location) of the workspace 60 of the first process on the timer list. This is illustrated in Figure 6.
  • the processes in the timer list are all linked in a time ordered manner.
  • Each process workspace contains a value in the time location 69 indicating the time at which the process may be scheduled.
  • the TLink location 68 of each process workspace includes a pointer to the 0 location of the workspace of the next process on the timer list.
  • Location 65 of each process workspace on the list stores a pointer to the next instruction in the program sequence 181 for use when the process is scheduled and becomes the current process.
  • the VALID TIME FLAG 84 is set to the value 1 when there are processes on the timer list and has the value 0 if there are no processes on the timer list.
  • the NEXT TIME REG 85 contains the time taken from location 69 of the process at the front of the timer list. In this way the register 85 contains an indication of the earliest time at which any of the processes on the associated timer list should be scheduled. No register is provided for the timer list to indicate the back of the list.
  • the workspace of the last process on the timer list has the special value "NotProcessp" in the TLINK location 68 of its workspace.
  • the front of the list is indicated by use of the memory location 90 rather than a register.
  • the front of the list is identified by use of a memory location in the same way as all intermediate entries on the list are identified and this simplifies the actions necessary to insert or delete further processes onto the timer list in a sequentially time ordered manner. It will be appreciated that it may become necessary to insert a process before the existing first process on the timer list or it may be necessary to insert it partway through the list depending on the time at which the process to be inserted is to be scheduled.
  • the timer logic shown in Figure 3 compares the times shown in the NEXT TIME REG 85 (indicating the first time for scheduling any of the processes on the list) with the time indicated by the CLOCK REG 81 and if the time for scheduling that first process has arrived, the timer logic provides an appropriate request signal to the condition multiplexor 36.
  • the processor responds to such request signals by removing the first process from the appropriate timer list and updating the appropriate VALID TIME FLAG, NEXT TIME REG and TPTR location 90.
  • a process may perform an instruction including "Timerlnput” by loading the time after which the process should be rescheduled into the AREG 54 and then executing the operation "TimerInput”. Firstly the processor checks whether the current time indicated by the CLOCK REG is after the time indicated by the AREG and if so no action occurs so that the process remains scheduled. If however this condition is not met the sequence specified in the definition of "TimerInput” occurs in that the special value "Waiting.p" is written into the STATE LOCATION 67 of the process workspace.
  • the time at which the process should be rescheduled is to be after that shown in the AREG and consequently the time indicated in the AREG is incremented by 1 to indicate the time at which the process should be rescheduled.
  • the processor carries out the procedure "InsertInTimerList" which writes into the time location 69 of the process workspace the time at which the process should be rescheduled and it causes the process to be fitted into the appropriate timer list at a position in that list such that the processes follow a time ordered sequence. It also sets the SNPFlag to a value 1 so that the processor starts executing another process.
  • the process which executed the "TimerInput" instruction will be rescheduled when an appropriate amount of time has passed.
  • alternative processes select one of the number of alternative components for execution.
  • Each component of the alternative consists of an input or a skip followed by a corresponding process.
  • the present example is able to execute a timer alternative process which selects one of a number of alternative components for execution.
  • Each component of the timer alternative may consist of a message channel input (from either an internal channel or an external channel), a skip or a timer input followed by a corresponding process.
  • a message channel input component may be selected if the channel is ready and a skip component may always be selected as described in the above referred to copending patent applications.
  • a timer input component may be selected when the value in the CLOCK REG is AFTER the time specified in the timer input.
  • the present example executes alternative processes which are not dependent on a timer input in precisely the same manner as has already been described in the above mentioned copending patent applications and that description will not be repeated in this specification.
  • each component is examined to determine if one or more of them can be selected. If no component can be selected, the process is descheduled until one of them can be selected. The process will then be rescheduled, the components reexamined and one of them selected.
  • the examination of message channel input components and skip components is performed as described in the above mentioned copending patent applications. When all components have been examined the state location 67 of the process workspace contains one of the two special values "Enabling.p" or "Ready.p".
  • the TLink and Time Locations 68 and 69 respectively are used for special purposes.
  • the TLink location 68 takes one of the two special values "TimeSet.p” or “TimeNotSet.p”. It is initialised to "TimeNotSet.p” indicating that no timer input has yet been examined and changes to "TimeSet.p” when the first timer input is examined.
  • the time location 69 is initialised to the time specified. Subsequently when each timer input is examined, the time location is updated to the time specified if that time is earlier than the time recorded in the time location 69.
  • the time location 69 holds the earliest time specified by any timer input.
  • the alternative process can select a timer input component if and only if the TLink location 68 contains the value "TimeSet.p" and the value of CLOCK REG is AFTER the time in the time location 69.
  • the Timer Alternative process determines if any component can be selected using the State, TLink and Time locations 67, 68 and 69. If no component can be selected the process is descheduled and if any timer input component has been examined, the process is placed on the appropriate timer list. When there is at least one component which can be selected each component is reexamined and the first selectable component is selected. As described in the above mentioned copending patent applications, the 0 location of the process workspace 60 is used to record which if any component has been selected. The reexamination of channel input components and skip components is performed as described in the above mentioned copending patent applications. The reexamination of the timer input components is as follows using the TLink and Time Locations 68 and 69.
  • Timer Alternative process is not on the timer list when the first timer input component is reexamined then either the process had been placed on the timer list and had subsequently been removed or the process had not been placed on the timer list at all.
  • the Time Location 69 contains the time at which the earlier timer input component became selectable.
  • the Time Location 69 contains the value of "CLOCK REG" immediately after examination of the component processes.
  • the Time Location retains the same value for all reexaminations of the timer input components.
  • a timer input component will be selectable if and only if the content of the time location 69 is AFTER the specified time.
  • Timer Alternative process is still on the timer list when the first timer input component is reexamined, there is no selectable timer input component but there must be a selectable channel input component.
  • the first reexamination of a timer input component removes the process from the timer list and sets the TLink location 68 to the value "TimeNotSet.p" preventing the selection of any timer input component. In this case no use is made of the Time Location 69.
  • timer alternative start followed by "enable timer” for each of the timer components.
  • the processor will also execute “enable channel” for each and every message channel if they are incorporated in the alternative construction. This is followed by “timer alternative wait” and then “disable timer” for each of the timer inputs and “disable channel” for any channel inputs. This is followed by the operation "Alternative End”.
  • the first instruction executed by a timer alternative process is the "timer alternative start” operation and as can be seen from the definition of that operation, in accordance with line 2 the special value "enabling p" is written into the state location 67 for the process and in accordance with line 3, the special value "TimeNotSet.p” is written into the TLink location 68 for the process workspace.
  • Any channel input components and skip components are examined by "enable channel” and “enable skip” operations as described in the above mentioned copending patent applications.
  • Any timer input components are examined by loading a guard value into the AREG and the specified time for the timer component into the BREG and then executing an "enable timer" operation.
  • lines 2 and 3 check whether the guard value in the AREG is false. If it is false the timer input component is to be ignored and the instruction has no other effect. Provided the guard value is not false in accordance with line 5 of the definition, the processor carries out the sequence beginning at line 7 of the definition.
  • Line 15 requires that the time value recorded in the time location 69 for the process is loaded into the OREG 57 and the value of this time is tested to see whether it meets the condition of line 17 of the definition. If that time is AFTER the time indicated in the BREG then the time in the BREG is written into the time location 69 for the process. Lines 19 and 20 indicate that if the time indicated in the OREG was not AFTER the time indicated in the BREG no action is taken. Finally the BREG is loaded with the value from the CREG as required by line 21 of the definition. In this way the process examines each of the possible timer inputs and the time location 69 of the process is updated so that after the examination it contains the time of the earliest timer component. It will therefore be seen that the succession of "enable timer" operations for each of the timer components effectively determines the earliest time of any of the components and progressively updates the time location 69 with the earliest time of any of the examined components.
  • the process then executes the operation "timer alternative wait".
  • this initialises the 0 location of the process workspace to the value -1 and then tests to determine if any component of the alternative process is already selectable.
  • it reads into the BREG the value from the TLink location 68 and reads into the AREG the value from the Time location 69.
  • Lines 5 and 6 require that if the process had the value "TimeSet.p" and the CLOCK REG shows a time AFTER the time indicated in the time location 69 of the process then the sequence defined in lines 8 and 9 occurs.
  • the special value "Ready.p” is written into the state location 67 for the process and the current time indicated by the CLOCK REG is written into the time location 69 for the process.
  • the process is not descheduled and may move onto its next instruction. If however the condition of line 6 of the definition was not true then the process moves to line 12 of the definition. It tests the contents of the state location 67 for the process by loading this into the CREG and line 14 tests whether this contains the value"Ready.p". If so then in accordance with line 16 the current time indicated by the CLOCK REG is written into the time location 69 for the process and the process is not descheduled. It is ready due to another of the alternative inputs and the process may move on to the next instruction.
  • the BREG may have the value "TimeNotSet.p" if the process is not awaiting any timer components and this will arise where the process is still awaiting a channel input rather than a timer input.
  • the sequence following line 25 occurs and the instruction pointer for the process is stored in the IPTR location 65 of the process workspace and the SNPFlag is set to the value 1 so that the process is descheduled.
  • the next instruction carried out by the process if it is not descheduled or when it is subsequently rescheduled, will be to effect the operation "disable timer" for each of the timer components, “disable skip” for any skip components and “disable channel” for any channel components.
  • the channel input components and skip components are reexamined by the "disable channel” and “disable skip” operations as described in the above mentioned copending patent applications.
  • the timer alternative process reexamines the timer input components in accordance with the definition of the operation "disable timer". Initially the AREG is loaded with a code offset to indicate the offset necessary in the program sequence in order to locate subsequent program instructions should that alternative component be selected by the process.
  • a guard value is loaded into the BREG and the CREG is loaded with the time at which the process is to be scheduled.
  • Line 2 of the definition checks whether the guard value is false and if so then this component cannot be selected and the AREG is loaded with the value MachineFALSE. Provided the guard was not false the process examines the content of the TLink location 68 for the process. There are three cases to consider. Firstly the TLink location may contain the value "TimeSet.p" in accordance with line 10 of the definition in this case the component is selectable if the time in the time location 69 is AFTER the specified time in the CREG. This is the condition in line 14 of the definition and if met then the process carries out the procedure "IsThisSelectedProcess".
  • the "disable timer” operation finds that the TLink location 68 of the process contains a value other than "TimeSet.p” or “TimeNotSet.p” this corresponds to the situation in line 18 of the definition of "disable timer". This will arise when the process is still on a timer list such that the TLink location 68 includes a pointer to a further process on the list. The timer component is therefore not selectable as the process is still waiting on a timer list and the process is removed from the timer list by the procedure "delete from timer list”. This causes the value "TimeNotSet.p" to be written into the TLink location 68 for the process.
  • the "disable timer” operation may find that the TLink location contains the value "TimeNotSet.p" in accordance with line 8 of the definition.
  • the TLink location 68 was set to this value by a previous "disable timer” operation which was executed while the process was on a timer list and so this component is not selectable. Consequently the AREG is set to the value MachineFALSE in accordance with line 9 of the definition.
  • the process carries out the operation "Alternative End" and in accordance with that definition, it first loads into the OREG the code offset which has been stored in the 0 location of the process workspace and then adjusts the pointer value in the IPTR REG by the offset indicated in the OREG. This causes the process to select the next instruction in a program sequence with an offset appropriate to the alternative process selected.
  • the AREG may be loaded with a value for example 14 indicating that the process wishes to continue when the clock register contains a value AFTER 14. If the instruction is executed at a time when the clock register contains the value 20, the processor will in accordance with the first two lines of the definition of "timer input” check whether the value in the clock register is after that indicated in the AREG. In this example that condition will apply and so the process will continue without descheduling the process.
  • Figure 9A shows the position immediately before execution of the "timer input” instructions.
  • the AREG 54 contains the value 30 indicating that the process wishes to be scheduled only when the time in the CLOCK REG 81 is AFTER 30.
  • the CLOCK REG currently contains the time value 20 and the valid time flag 84 is set to 0 indicating that there are no processes on the priority 1 timer list.
  • the process must be descheduled.
  • the special value "waiting.p” is written into the state location 67 of the process X and the value in the AREG is incremented so that it contains the time at which the process should be scheduled.
  • the process is then inserted into the timer list and the position is as shown in Figure 9B.
  • the CLOCK REG has now incremented to 22.
  • the valid time flag 84 is now set to the value 1 indicating that there is at least one process on the timer list.
  • the NEXT TIME REG 85 contains the value 31 which is the time at which the first process on the timer list should be scheduled and the TPTR location 90 contains the workspace pointer of process X being the first (and only) process on the timer list.
  • the workspace of process X contains its instruction pointer (IPTR) in location 65, the special value "waiting.p” in location 67, special value "not process.p” in location 68 indicating that this is the last process on the timer list and the value 31 in location 69 indicating the time at which the process can be rescheduled.
  • This example illustrates how the insert flag 82 is used to cause a process to be inserted into the timer list at the correct position in a time ordered sequence.
  • a process P performs a timer input operation which causes it to be descheduled. It is assumed that process P is a priority 1 process and is the only process executing. It is further assumed that there are three other processes waiting on the priority 1 timer list, these are process X waiting for time 25, process Y waiting for time 26 and process Z waiting for time 29.
  • Figure 10A illustrates the position just before executing the timer input instruction. Process P is executing and the AREG contains the time 27. The CLOCK REG contains the time 20.
  • the valid time flag is set to the value 1 indicating that the timer list is in use and the NEXT TIME REG contains the value 25 indicating that the time associated with the earliest process on the timer list is 25. It can be seen that there are three processes on the timer list.
  • the TPTR location 90 contains a pointer to the first of these which is process X.
  • the TLink location 68 of process X contains a pointer to the second process Y which in turn contains a pointer to the third process Z.
  • the TLink location 68 of process Z contains a special value "Not Process.p" indicating that process Z is the last process on the timer list. It can be seen that the timer list is ordered with the earlier process first and the latest process last.
  • the next action that is performed by the processor is the procedure "insert step”.
  • this procedure causes the TREG 49 to be loaded with the time associated with process X (that is 25) and will compare that with the time associated with process P (that is 28). Since 28 is AFTER 25 the processor has not yet found the correct place to insert process P into the timer list and the "insert step” procedure causes the BREG to be set to a pointer to the TLink location 68 of process X and the CREG is set to the contents of that location. The procedure then terminates leaving the insert flag set. The resulting situation is shown in Figure 10C. The next action of the process will be to perform the procedure "insert step" again. This will be executed in a similar manner to that previously described and will result in the situation shown in Figure 10D.
  • the processor clears the insert flag and inserts process P into the timer list between process Y and process Z by writing the workspace pointer of process P into the TLink location 68 of process Y and writing the workspace pointer of process Z into the TLink location 68 of process P.
  • the processor then resets the NEXT TIME REG 85 to the time associated with the first process on the timer list and sets the valid time flag to the value 1.
  • the processor writes the instruction pointer of process P into the IPTR location 65 of process P and sets the SNPFlag 58 to the value 1 to cause process P to be descheduled as the next action of the processor.
  • the resulting situation is shown in Figure 10E.
  • FIG. 11A shows the position immediately after executing the "timer alternative start” instruction.
  • the state location 67 contains the special value "enabling.p” and the TLink location 68 contains the special value "TimeNotSet.p".
  • the AREG will contain the value MachineTRUE and the BREG will contain the time associated with this timer input which is 26.
  • the processor When the enable timer instruction is executed the processor reads the TLink location 68 and finds that it contains the value "TimeNotSet.p” indicating that no timer input component has previously been examined. The processor therefore sets the TLink location 68 to the special value "TimeSet.p” and the time location 69 to the value 26. This is the position shown in Figure 11B. Immediately before the second "enable timer” instruction is executed the AREG will contain the value MachineTRUE and the BREG will contain the value 25 being the time associated with the second timer input component. When the enable timer instruction is executed the process reads the TLink location 68 and finds that it contains the value "TimeSet.p” indicating that the time location contains the earliest time associated with any previous timer input component. The processor therefore reads the time location 69 and determines that the time specified for this component which is 25, is earlier than the time read from the time location which contains the value 26. The processor therefore writes the new value 25 into the time location and the position is as shown in Figure 11C.
  • FIG. 12A to 12C illustrates a timer alternative process P with two timer input components where the process P is not descheduled. It is assumed that process P is the only runnable process, that it is a priority 1 process and that the time specified in the first timer input component is 26 and the time of the second timer input component is 25. Execution of the "timer alternative start” instruction and the examination of the timer input components is as previously described in Example 4 and the situation immediately before executing the "timer alternative wait” instruction is as shown in Figure 11C. The first action of the "timer alternative wait” is to write the value -1 into the 0 location of the workspace 60 of the process P. This is the location used to select a component from a plurality of alternatives.
  • the processor next determines that process P can continue without descheduling as the time in the CLOCK REG is after the time in the time location 69.
  • the processor therefore writes the special value "Ready.p” into the state location 67 and the value of the clock register is written into the time location 69.
  • the position just before the first "disable timer" instruction is illustrated in Figure 12B.
  • the AREG contains the offset from the "Alternative End” instruction to the sequence of instructions in the program associated with the first timer input component, the BREG contains the value MachineTRUE and the CREG contains the time associated with this timer component which is 26.
  • the process then executes the "disable timer" instruction which reads the TLink location 68 and determines that it contains the value "TimeSet.p". Consequently it reads the value 30 from the time location and as 30 is AFTER 26 this timer input component is selectable.
  • the processor then performs the procedure "IsThisSelectedProcess" which will select this component as the 0 location of the process workspace still contains the value -1. The resulting situation is shown in Figure 12C.
  • the second timer input component cannot now be selected and when the Alternative End instruction is executed the workspace for process P will still be as illustrated in Figure 12C.
  • FIG. 13A to 13F shows a timer alternative process P with two timer input components where the process P is descheduled. It is assumed that process P is the only runnable process, that the process has priority 1, the time specified in the first timer input component is 26, the time specified in the second timer input is 25 and there are no processes on the timer list.
  • the execution of the "timer alternative start” instruction and the examination of the timer input components is as previously described in Example 4 and the position immediately before executing "timer alternative wait” instruction is as previously shown in Figure 11C.
  • the first action of the "timer alternative wait” instruction is to write the value -1 into the zero location of the workspace of process P.
  • the processor compares the value in the time location 69 with the value of the clock register and, finding that the process cannot proceed due to a timer input, checks the state location 67 of the process. As this contains "enabling.p" the process is placed on the timer list and descheduled. This is the position shown in Figure 13A. The valid time flag is set to the value 1 indicating that the timer list is not empty.
  • the NEXT TIME REG contains the value 26 which is the time at which process P will become ready to execute.
  • the TPTR location 90 contains a pointer to the workspace of the process P and the TLink location 68 of process P contains the special value "not process.p" indicating that it is the last process on the list.
  • FIG. 14A This illustrates a timer alternative process P with one timer input component and one message channel input component. It is assumed that process P is the only runnable process, that it has priority 1, and a specified time of 40. There are no processes on the timer list and the channel referred to by the channel input component is initially "Ready” and the timer input component is not selectable.
  • FIGs 14A to 14D This example is illustrated in Figures 14A to 14D.
  • the process P executes a "timer alternative start” instruction, loads its registers appropriately and executes an "enable timer” instruction. The process then loads the registers in preparation for a "enable channel” instruction. This results in the situation shown in Figure 14A. As the channel is "Ready” the situation after execution of the "enable channel” instruction is as shown in Figure 14B.
  • the process then executes a "timer alternative wait” instruction.
  • the time in the CLOCK REG has the value 11 which is not AFTER the time 40 indicated in the time location 69 for the process P. Therefore the processor checks the state location 67 which contains the value "Ready.p” and consequently writes into the time location 69 the time value in the clock register.
  • the situation on completion of the "timer alternative wait” instruction is as shown in Figure 14C.
  • the situation immediately before the "disable timer” instruction is executed is as shown in Figure 14D.
  • the timer input component will not be selected because the time value 12 in the time location 69 is not AFTER the time associated with the component.
  • the process will then execute a "disable channel” instruction which will select the channel input component.
  • the processor first executes a "timer alternative start” instruction, an "enable timer” operation for the one timer input component and an “enable channel” operation on the channel 70. The position is then as shown in Figure 15A.
  • the process P is not yet descheduled, the "state" location 67 has been initialised to "enabling.p” to indicate that the process is carrying out an alternative input.
  • the TLink location 68 has been set to the value "TimeSet.p” indicating that a timer input has been examined.
  • the "Time” location 69 has been set to the earliest time of any timer input examined which in this case is 40 being the only timer input examined.
  • the timer list has two descheduled processes X and Y with scheduling times of 35 and 55 respectively.
  • the processor then executes a "timer alternative wait" instruction for process P.
  • the process then executes a "disable timer” operation and this reads the "TLink location" 68 for process P and determines that the process is still on a timer list as it contains the workspace pointer to the next process on the timer list. As process P is still on the list the time for that timer input component has not arrived and consequently the timer component is not selectable.
  • the AREG is therefore set to MachineFALSE and the procedure "delete from timer list" is performed. This sets the DELETE FLAG to the value 1 loads, the BREG with a pointer to the TPTR location 90 and loads the CREG with the contents of the TPTR location 90.
  • the instruction then terminates leaving the position as shown in Figure 15D.
  • the next action of the processor is to perform the procedure "delete step".
  • the TPTR location contains the workspace pointer of process X, it will be the workspace pointer of process X which is first loaded into the CREG and consequently in carrying out the procedure "delete step” the condition of line 2 of the definition of "delete step” will apply in that the CREG does not contain the workspace pointer of process P.
  • process P is removed from the timer list by loading into the CREG the value currently held in the TLink location 68 for process P (that is a pointer to the workspace pointer of process Y) and then writing the value from the CREG into the location indicated by the BREG which is the TLink location for process X.
  • the contents of the TLink location of process X are changed to replace the pointer to the workspace of process P by a pointer to the workspace of process Y.
  • the processor checks whether there are any processes left on the timer queue by following line 13 of the "delete step” procedure in which the BREG is loaded with the contents of the TPTR location. If in accordance with line 15 of the definition, this has the value "not process.P” then there are no processes left on the list. The valid time flag is then set to zero in accordance with line 17. If, on the other hand, a value other than "not process.P" was found in accordance with line 18 of the definition then there is a further process on the timer list and the NEXT TIME REG is updated by taking the time from the Time location 69 of the process indicated by the BREG in accordance with line 20 of the definition.
  • the appropriate code offset will be loaded into the 0 location of the workspace for process P so that on completing the next instruction "Alternative End" the code offset will be added to the instruction pointer for process P so that the process moves to the correct part of the program in accordance with the selection of the channel input.
  • This example program is arranged to calculate the number of revolutions made per second by a flywheel.
  • the process is arranged to communicate through two channels one called “rotation” and the other called “rps” which represents revolutions per second.
  • the process is arranged to input from the channel "rotation” whenever the flywheel completes a revolution.
  • the process may also receive a timer input so that the process may respond to the occurrence of a predetermined time.
  • the predetermined time is the successive passage of one second intervals.
  • the process is arranged to output each second through the channel "rps" the number of revolutions which have occurred during that second.
  • the following additional notation is used:-
  • the current value of the processor's clock is represented NOW.
  • a "timer" input is represented as
  • This input specifies that the process may not.proceed until the processor clock holds a time AFTER the time t.
  • Line 1 of the program specifies that the process uses two variables one of which is called “Rotations” which is used to count the number of rotations occurring in a one second interval and the other variable “EndOfInterval” is used to record the value of the processor's clock which will indicate the termination of the current one second interval.
  • Line 2 specifies that a sequence is to be followed as set out in lines 3 to 6.
  • the count of number of rotations is set to 0.
  • line 4 the current value of the processor's clock is read so that line 5 can calculate the value of the processor's clock for the end of the one second interval.
  • the value 10000 used in line 5 is the number of times the processor's clock increments in one second.
  • Line 6 indicates that the alternative process which follows between lines 7 and 14 is to be repeated continuously.
  • Line 7 identifies the process as a timer alternative process.
  • Lines 8 and 10 set out the two alternative inputs.
  • Line 8 may input a signal from the channel "Rotation” if the flywheel has completed a rotation. If this input is selected then the corresponding process on line 9 is executed which increments the number of rotations counted in the current one second interval.
  • the timer input on line 10 can be selected when the current one second interval has been completed. If this timer input to the process is selected then the corresponding process on lines 12, 13 and 14 will be executed.
  • Line 12 provides an output through the channel "rsp" indicating the count of the number of rotations which have occurred in the one interval.
  • Line 13 resets the rotation counter to 0 and line 14 calculates the time of the end of the next one second period.
  • lines 1 and 2 initialise the count of the numb er of rotations to 0.
  • Lines 3 and 4 use a pfix function in order to operate load timer to read the processor clock.
  • Lines 6 to 11 use successive pfix functions and an add constant function to calculate the value of the processor's clock at the end of a one second interval.
  • the timer alternative input begins at line 13, lines 13 and 14 use the pfix function in order to operate "timer alternative start”.
  • Line 15 loads a pointer to the channel "Rotation” and lines 16a and 17 use the pfix function to operate "enable channel”.
  • Line 18 loasds the value of the variable "EndOfInterval".
  • Line 19 loads the guard value and lines 20 and 21 use a pfix function to operate "enable timer".
  • Lines 22 and 23 carry out "timer alternative wait”.
  • Lines 24 to 27 reexamine the channel input.
  • Line 24 identifies the channel "Rotation”.
  • Line 25 loads the guard value MachineTRUE.
  • Line 26 loads the instruction offset which will be necessary if the channel input is selected.
  • Lines 26a and 27 carry out the operation "disable channel”.
  • Lines 28 to 32 reexamine the timer input.
  • Line 28 loads the variable "EndOfInterval”.
  • Line 29 loads a guard value
  • Line 30 loads the instruction offset which will be necessary if the process selects the timer input and lines 31 and 32 carry out "disable timer”.
  • Lines 32a and 33 carry out "Alternative End”.
  • Line 35 is the first instruction which will be executed if the channel input is selected.
  • Line 45 is the first instruction which will be executed if the timer input is selected.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Multi Processors (AREA)
  • Executing Machine-Instructions (AREA)
  • Computer And Data Communications (AREA)
EP84308344A 1984-11-30 1984-11-30 Mikrorechner für zeitabhängige Prozesse Expired - Lifetime EP0183877B1 (de)

Priority Applications (3)

Application Number Priority Date Filing Date Title
DE8484308344T DE3483007D1 (de) 1984-11-30 1984-11-30 Mikrorechner fuer zeitabhaengige prozesse.
EP84308344A EP0183877B1 (de) 1984-11-30 1984-11-30 Mikrorechner für zeitabhängige Prozesse
KR1019850006994A KR930009757B1 (ko) 1984-11-30 1985-09-24 시간 의존 처리용 마이크로컴퓨터

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
EP84308344A EP0183877B1 (de) 1984-11-30 1984-11-30 Mikrorechner für zeitabhängige Prozesse

Publications (2)

Publication Number Publication Date
EP0183877A1 true EP0183877A1 (de) 1986-06-11
EP0183877B1 EP0183877B1 (de) 1990-08-16

Family

ID=8192828

Family Applications (1)

Application Number Title Priority Date Filing Date
EP84308344A Expired - Lifetime EP0183877B1 (de) 1984-11-30 1984-11-30 Mikrorechner für zeitabhängige Prozesse

Country Status (3)

Country Link
EP (1) EP0183877B1 (de)
KR (1) KR930009757B1 (de)
DE (1) DE3483007D1 (de)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0415515A2 (de) * 1989-09-01 1991-03-06 Litton Systems, Inc. Rechnersystem
WO1999061985A1 (en) * 1998-05-26 1999-12-02 The Dow Chemical Company Distributed computing environment using real-time scheduling logic and time deterministic architecture
US6748451B2 (en) 1998-05-26 2004-06-08 Dow Global Technologies Inc. Distributed computing environment using real-time scheduling logic and time deterministic architecture

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2030331A (en) * 1978-01-24 1980-04-02 Plessey Co Ltd Real-time Data Processing System for Processing Time Period Commands

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2030331A (en) * 1978-01-24 1980-04-02 Plessey Co Ltd Real-time Data Processing System for Processing Time Period Commands

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
PROCEEDINGS IFIP CONGRES, Stockholm, 5th-10th August 1974, pages 257-260, North-Holland Publishing Company, Amsterdam, NL; E. FIZZAROTTI: "A communications oriented multitasking system" *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0415515A2 (de) * 1989-09-01 1991-03-06 Litton Systems, Inc. Rechnersystem
EP0415515A3 (en) * 1989-09-01 1992-02-19 Litton Systems, Inc. Computer system
WO1999061985A1 (en) * 1998-05-26 1999-12-02 The Dow Chemical Company Distributed computing environment using real-time scheduling logic and time deterministic architecture
US6748451B2 (en) 1998-05-26 2004-06-08 Dow Global Technologies Inc. Distributed computing environment using real-time scheduling logic and time deterministic architecture

Also Published As

Publication number Publication date
EP0183877B1 (de) 1990-08-16
KR930009757B1 (ko) 1993-10-09
KR860004351A (ko) 1986-06-20
DE3483007D1 (de) 1990-09-20

Similar Documents

Publication Publication Date Title
US4989133A (en) System for executing, scheduling, and selectively linking time dependent processes based upon scheduling time thereof
EP0145244B1 (de) Mikrorechner
EP0061096B1 (de) Datenverarbeitungssystem für parallele Verarbeitung
US4675806A (en) Data processing unit utilizing data flow ordered execution
US4320455A (en) Queue structure for a data processing system
US4468736A (en) Mechanism for creating dependency free code for multiple processing elements
US3825902A (en) Interlevel communication in multilevel priority interrupt system
US4466061A (en) Concurrent processing elements for using dependency free code
JPS5911943B2 (ja) デ−タ処理装置の為のトラツプ機構
US3654621A (en) Information processing system having means for dynamic memory address preparation
EP0125044B1 (de) Mikrocomputer mit Interprozesskommunikation
AU603876B2 (en) Multiple i/o bus virtual broadcast of programmed i/o instructions
CA2056356C (en) Interruption handling system
EP0183877B1 (de) Mikrorechner für zeitabhängige Prozesse
JPS59106075A (ja) デ−タ処理システム
WO1986003311A1 (en) Microcomputer for time dependent processes
EP0105125A2 (de) Datenverarbeitungssystem
CA1302580C (en) Apparatus and method for using lockout for synchronization of access to main memory signal groups in a multiprocessor data processing system
EP0206335A2 (de) Unterbrechungsmethode für ein Datenverarbeitungssystem
EP0057312A2 (de) Unterprogramm-Steuerungsschaltung
KR930005767B1 (ko) 마이크로프로그램 제어를 기초로한 데이타 처리장치
EP0177268A2 (de) Programmierbare Datenpfadbreite in einer programmierten Einheit mit mehreren Niveaus von Unterbefehlssätzen
US4467410A (en) Multi-phase subroutine control circuitry
JPH08137690A (ja) プログラム実行制御方法
KR920005298B1 (ko) 마이크로 컴퓨터

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): DE FR GB IT NL SE

17P Request for examination filed

Effective date: 19861209

17Q First examination report despatched

Effective date: 19880824

RAP3 Party data changed (applicant data changed or rights of an application transferred)

Owner name: INMOS LIMITED

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

ITF It: translation for a ep patent filed

Owner name: STUDIO INGG. FISCHETTI & WEBER

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE FR GB IT NL SE

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SE

Free format text: THE PATENT HAS BEEN ANNULLED BY A DECISION OF A NATIONAL AUTHORITY

Effective date: 19900816

REF Corresponds to:

Ref document number: 3483007

Country of ref document: DE

Date of ref document: 19900920

ET Fr: translation filed
PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: NL

Effective date: 19910601

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

NLV4 Nl: lapsed or anulled due to non-payment of the annual fee
26N No opposition filed
ITTA It: last paid annual fee
REG Reference to a national code

Ref country code: FR

Ref legal event code: TP

REG Reference to a national code

Ref country code: GB

Ref legal event code: 732E

REG Reference to a national code

Ref country code: FR

Ref legal event code: D6

REG Reference to a national code

Ref country code: GB

Ref legal event code: IF02

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20021108

Year of fee payment: 19

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20021127

Year of fee payment: 19

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20021205

Year of fee payment: 19

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20031130

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20040602

GBPC Gb: european patent ceased through non-payment of renewal fee

Effective date: 20031130

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: FR

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20040730

REG Reference to a national code

Ref country code: FR

Ref legal event code: ST